分布式并行混合遗传算法在无功优化中的应用  

Application of Distributed Parallel Hybrid Genetic Algorithm in Reactive Power Optimization

作　　者：李运华[1] 吴宏昺[1] 盛万兴[2] 刘科研[1] 

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,北京100083 [2]中国电力科学研究院,北京100085

出　　处：《电力系统及其自动化学报》2008年第2期36-41,共6页Proceedings of the CSU-EPSA

基　　金：973专业基金(G1998030416/05)

摘　　要：将模拟退火算法与普通遗传算法结合,同时引入分布式MPI技术,实现了分布式的多种群算法。该算法用个体迁移策略来协调优化各个种群,用最优值恢复策略来稳定优化方向,并通过分布式计算的应用,有效地缩短了计算时间。通过运用IEEE14节点标准数据运算,证明了这种算法在电力系统无功优化计算中具有很高的稳定性和快速求解能力。Based on the combination of normal genetic algorithm with simulate-annealing algorithm and the introduction of distributed MPI technique,a distributed multi-colony algorithm is proposed. This algorithm coordinates the optimization of all the courses by individual transfer ,and ensures the stability of optimization direction by the strategy of oriental resumption. Computation time is sharply reduced by the application of distributed technique. Case study on IEEE 14-node system shows that the proposed algorithm can solve the problem of reactive power optimization with high stability and speed.

关 键 词：无功优化 消息传递界面 分布式并行混合遗传算法 模拟退火 个体迁移 最优值恢复 

分 类 号：TM744[电气工程—电力系统及自动化]